This thesis investigates the representation of operational decision makers within simulation modelling. Artificial Intelligence concepts, such as expert systems focus on the problem of representing, in high-level code, complex real-world decision making problems. The author therefore proposes that the use of expert system technology may provide an improved means of representing operational decision tasks and that as a consequence, apriori possibilities may exist in the context of model experimentation based on alternative operational policies. The thesis further investigates the nature of operational decision making and the potential need to represent within a model, inter-dependencies between decision makers. A prototype system called ESSIM is developed which comprises of two interlinked components, a discrete event simulation module and expert system module. The benefits of the proposed approach are then assessed by comparing the functionally of ESSIM with conventional modelling techniques. The comparison is carried out by developing three alternative models of an automated container port, one of these using ESSIM. Experiments were then devised and executed which seek to draw conclusions on the thesis proposal.

The increasing waiting times to access healthcare services are a major concern for pa-tients in hospitals. Due to the unpredictability of health issues, hospitals and clinical ser-vices are provided to patients even without prescheduled medical appointments. Unex-pected and random patient arrivals can result in high waiting times. Waiting occurs most-ly because of insufficient resources available compared to demanding service delivery requirements at a given time. Thus, appropriate management of resource scheduling over time can help reduce patient wait times. So far, simulation has mostly been used as a support for strategic decision making in healthcare environments. We are proposing a complementary approach, namely, real-time simulation, to support operational decision making rather than long-term strategic decision making. Real-time simulation is a technique used to get a timely prediction of the system status in a near future (e.g., a few hours). Hospitals can benefit from the capa-bilities of real-time simulations by predicting upcoming bottleneck occurrences in patient care processes and make effective decisions in the present time to avoid undesirable out-comes in the near future. This research presents real-time simulation capabilities for short-term operational decision making of patient care processes in hospitals and the possible ways to run alter-native scenarios and evaluate their results to come up with the most effective solution considering various factors. This thesis also provides tool support based on a leading simulation environment, namely Arena. The tool-supported methodology is evaluated through a realistic cardiac care process in an Ontario community hospital, with encourag-ing results.

The increasing waiting times to access healthcare services are a major concern for pa-tients in hospitals. Due to the unpredictability of health issues, hospitals and clinical ser-vices are provided to patients even without prescheduled medical appointments. Unex-pected and random patient arrivals can result in high waiting times. Waiting occurs most-ly because of insufficient resources available compared to demanding service delivery requirements at a given time. Thus, appropriate management of resource scheduling over time can help reduce patient wait times. So far, simulation has mostly been used as a support for strategic decision making in healthcare environments. We are proposing a complementary approach, namely, real-time simulation, to support operational decision making rather than long-term strategic decision making. Real-time simulation is a technique used to get a timely prediction of the system status in a near future (e.g., a few hours). Hospitals can benefit from the capa-bilities of real-time simulations by predicting upcoming bottleneck occurrences in patient care processes and make effective decisions in the present time to avoid undesirable out-comes in the near future. This research presents real-time simulation capabilities for short-term operational decision making of patient care processes in hospitals and the possible ways to run alter-native scenarios and evaluate their results to come up with the most effective solution considering various factors. This thesis also provides tool support based on a leading simulation environment, namely Arena. The tool-supported methodology is evaluated through a realistic cardiac care process in an Ontario community hospital, with encourag-ing results.

The concept of Operational decision making is not new, but in the context of Enterprise Resource Planning implementation as is the case at PFK Electronics (PTY) Ltd., it becomes part of what ERP process that makes organisations more competitive and profitable as Wallace and Kremzar (2001) comment, "Effective forecasting, planning, and scheduling is fundamental to productivity and ERP is a fundamental way to achieve it. Properly implementing ERP will give you a competitive advantage and help you run your business more effectively, efficiently and responsively." The Operational decision making process that forms part of the research question is a process known as Sales and Operational Planning process (SOP's). The research paper is looking at how Real Time Shop floor information can impact on the decision making process, as Wallace and Kremzar (2001 : 165) state that this is the most important element in making ERP systems work. The study looks at what information is required to make effective decisions in the SOP's process and how the respondents perceive information quality, its value and the impact on decision making in the SOP's process. The research sought to further investigate a diverse demographic population, the behavioural culture within the organisation and their perception on having this information in real time directly from the shop floor. Data was collected via a survey questionnaire and an interview process which also required the review of organisational documentation where available. The research aimed to investigate what information Operations (SOP's) need to make effective decisions to meet the organisational strategic objectives and to determine the best means of obtaining and communicating the information to managers. Developing this further the research was then broken down into three objectives: • To investigate 'what' information sales and operations managers need in order to make effective decisions that will have an impact on them meeting strategic performance objectives. • To evaluate 'how' real time shop floor information can improve the decision making process in Operations in meeting strategic objectives; by understanding the types and methods used in decision making and the significance of time on information. • To recommend suitable Information Communication Technology (ICT) systems to obtain and communicate information to managers effectively within operations. During this process time constraints were of a concern, which manifested in objective three not being able to be completed to the satisfaction of the researcher. However the research process, which included the survey and interviews of respondents, the researcher felt that respondents where honest and open in expressing their views which adds to the validity and reliability of the research. The results showed that there was a clear understanding of the topic amongst the respondents. It showed that the respondents had a good understanding of information and its importance in meeting strategic objectives by making better decisions. It became clear that ERP was generally working well but that timeliness of information was the biggest obstacle in meeting strategic objectives in balancing supply and demand. The research identified the information necessary for the SOP's process to be effective and it also showed the impact of having real time shop floor information available to the decision making process. It also showed that there would be other benefits as performance could also be improved by having real time performance indicators that are used to set the pace and common goal. The report also identified some cultural differences between PFK electronics culture and NUMSA unionised employees culture and the impact it has on relationships and communication. The report concludes with the research question being answered and with a recommendation that the single most important means of improving operational effectiveness is going to be to improve the quality of information used in decision making. The research has shown that getting this information in 'real time' or as close as possible to real time is the best means of improving the quality of information and its impact on management decisions. It is further recommended to investigate the implementation of systems such as Manufacturing Execution Systems to link the shop floor directly into the current ERP system and lastly to find a means of bridging the cultural differences between NUMSA unionised employees and PFK behavioural culture. / Graduate School of Business Leadership / M.B.A.

The concept of Operational decision making is not new, but in the context of Enterprise Resource Planning implementation as is the case at PFK Electronics (PTY) Ltd., it becomes part of what ERP process that makes organisations more competitive and profitable as Wallace and Kremzar (2001) comment, "Effective forecasting, planning, and scheduling is fundamental to productivity and ERP is a fundamental way to achieve it. Properly implementing ERP will give you a competitive advantage and help you run your business more effectively, efficiently and responsively." The Operational decision making process that forms part of the research question is a process known as Sales and Operational Planning process (SOP's). The research paper is looking at how Real Time Shop floor information can impact on the decision making process, as Wallace and Kremzar (2001 : 165) state that this is the most important element in making ERP systems work. The study looks at what information is required to make effective decisions in the SOP's process and how the respondents perceive information quality, its value and the impact on decision making in the SOP's process. The research sought to further investigate a diverse demographic population, the behavioural culture within the organisation and their perception on having this information in real time directly from the shop floor. Data was collected via a survey questionnaire and an interview process which also required the review of organisational documentation where available. The research aimed to investigate what information Operations (SOP's) need to make effective decisions to meet the organisational strategic objectives and to determine the best means of obtaining and communicating the information to managers. Developing this further the research was then broken down into three objectives: • To investigate 'what' information sales and operations managers need in order to make effective decisions that will have an impact on them meeting strategic performance objectives. • To evaluate 'how' real time shop floor information can improve the decision making process in Operations in meeting strategic objectives; by understanding the types and methods used in decision making and the significance of time on information. • To recommend suitable Information Communication Technology (ICT) systems to obtain and communicate information to managers effectively within operations. During this process time constraints were of a concern, which manifested in objective three not being able to be completed to the satisfaction of the researcher. However the research process, which included the survey and interviews of respondents, the researcher felt that respondents where honest and open in expressing their views which adds to the validity and reliability of the research. The results showed that there was a clear understanding of the topic amongst the respondents. It showed that the respondents had a good understanding of information and its importance in meeting strategic objectives by making better decisions. It became clear that ERP was generally working well but that timeliness of information was the biggest obstacle in meeting strategic objectives in balancing supply and demand. The research identified the information necessary for the SOP's process to be effective and it also showed the impact of having real time shop floor information available to the decision making process. It also showed that there would be other benefits as performance could also be improved by having real time performance indicators that are used to set the pace and common goal. The report also identified some cultural differences between PFK electronics culture and NUMSA unionised employees culture and the impact it has on relationships and communication. The report concludes with the research question being answered and with a recommendation that the single most important means of improving operational effectiveness is going to be to improve the quality of information used in decision making. The research has shown that getting this information in 'real time' or as close as possible to real time is the best means of improving the quality of information and its impact on management decisions. It is further recommended to investigate the implementation of systems such as Manufacturing Execution Systems to link the shop floor directly into the current ERP system and lastly to find a means of bridging the cultural differences between NUMSA unionised employees and PFK behavioural culture. / Graduate School of Business Leadership / M.B.A.

Syftet med denna studie är att undersöka till vilken utsträckning som företag inomSveriges företagsvärld inser nyttan och användbarheten med att använda sig utav ettaffärssystem vid beslut gällande verksamheten styrning. Uppsatsen avser även attundersöka sambandet mellan affärssystem och beslutstöd genom att undersökakopplingen mellan tydligt ställda mål på affärssystemet och sedermera användandet utavdet som beslutstöd. Studien är genomförd med en positivistisk ansats av kvantitativ art,då vi genom en enkätundersökning har undersökt hur 34 stycken företag ser på planeringoch det dagliga användandet av affärssystem. Empirin har sedan analyserats utifrån redantidigare befintliga teorier som vi även utgått ifrån vid framställningen av enkäten.Affärssystemen har i allt snabbare takt börjat utvecklas och dessutom blivit en allt mercentral del utav den ekonomiska styrningen för verksamheten. Målen för användare avaffärssystem inriktas dock främst mot att förbättra hanteringen utav transaktioner genomen standardisering utav affärsprocesser och integration mellan verksamhet och data(Holsapple & Sena, 2003). Men Davenport (1998) menar att behovet av att göra sundaaffärsbeslut är en utav de främsta anledningarna för att anskaffa ett affärssystem. Enstudie av Löfgren och Ericsson (2004) påvisar att svenska företag har svårt att få framinformation genom sitt affärssystem alternativt inte vet hur man skall användainformationen som erbjuds. Dem menar också att målen med systemlösning inomekonomifunktionen måste överensstämma med den övergripande styrfilosofin för attaffärssystemet ska kunna maximeras. Ytterligare en studie gjord av Holsapple & Sena(2005) visar på att det finns en viss, men tydlig, användning av affärssystem sombeslutstöd i den amerikanska företagsvärlden. Men vad skulle en liknande undersökning,utförd i Sveriges företagsvärld, ge för resultat? Kommer resultatet i så fall stärka det somLöfgren & Ericsson (2004) påstår gällande sambandet mellan mål med affärssystemetoch användandet av det som beslutsstöd?Med studien har vi kommit fram till följande slutsats. Företag i Sverige inser till viss delnyttan med att använda affärssystem som beslutsstöd, de lägger en betydlig vikt vid attplanera affärssystem både vad gäller beslutsstöd men främst i övriga frågor gällande attintegrera och standardisera verksamheten. Även i det dagliga användandet avaffärssystemet påträffas situationer, gällande beslutstöd, av betydelse, även om företagenlångt ifrån använder systemet till fullo. Vid jämförelsen med amerikanska företag ser viatt även de lägger stor vikt vid att planera affärssystem men att de desto oftare använderaffärssystemet för olika situationer i den dagliga verksamheten, vilket är en markantskillnad mot de svenska företagen.

A two level hierarchical simulation and decision modeling framework is proposed for electric power networks involving PV based solar generators, various storage, and grid connection. The high level model, from a utility company perspective, concerns operational decision making and defining regulations for customers for a reduced cost and enhanced reliability. The lower level model concerns changes in power quality and changes in demand behavior caused by customers' response to operational decisions and regulations made by the utility company at the high level. The higher level simulation is based on system dynamics and agent-based modeling while the lower level simulation is based on agent-based modeling and circuit-level continuous time modeling. The proposed two level model incorporates a simulation based optimization engine that is a combination of three meta-heuristics including Scatter Search, Tabu Search, and Neural Networks for finding optimum operational decision making. In addition, a reinforcement learning algorithm that uses Markov decision process tools is also used to generate decision policies. An integration and coordination framework is developed, which details the sequence, frequency, and types of interactions between two models. The proposed framework is demonstrated with several case studies with real-time or historical for solar insolation, storage units, demand profiles, and price of electricity of grid (i.e., avoided cost). Challenges that are addressed in case studies and applications include 1) finding a best policy, optimum price and regulation for a utility company while keeping the customers electricity quality within the accepted range, 2) capacity planning of electricity systems with PV generators, storage systems, and grid, and 3) finding the optimum threshold price that is used to decide how much energy should be bought from sold to grid to minimize the cost. Mathematical formulations, and simulation and decision modeling methodologies are presented. A grid-storage analysis is performed for arbitrage, to explore if in future it is going to be beneficial to use storage systems along with grid, with future technological improvement in storage and increasing cost of electrical energy. An information model is discussed that facilitates interoperability of different applications in the proposed hierarchical simulation and decision environment for energy systems.

Afin d’expliquer la diversité des pratiques des agriculteurs, cette thèse explore le rôle des caractéristiques du processus de décision dans les choix opérationnels réalisés. Sur la base d’une méthode à deux volets (modélisation et enquêtes), nous avons proposé un modèle conceptuel du processus de prise de décision des agriculteurs, caractérisé les différences possibles au sein de ce processus (appelées modes de décisions), mis en évidence le lien entre ces modes et les pratiques, précisé les modalités de simplification du processus de décision et identifié de nouveaux leviers d’action pour l’accompagnement des agriculteurs. De nature générique, les modèles et les méthodes d’enquêtes proposés ont été appliqués à un cas d’étude qui à permis d’expliquer l’hétérogénéité des pratiques antifongiques des agriculteurs sur le blé tendre en MidiPyrénées par leurs caractéristiques personnelles, par les caractéristiques de leurs exploitations et par leurs modes de décision / In order to explain the diversity of farmers’ practices, this these explores the decision-making process features involved in operational choices. Using a modeling and survey-based approach, we built a conceptual model of the farmer’s decision-making process, distinguished different use of the decision-making process (referred to as decision modes), highlighted the link between thesedecision modes and farmers’ practices, showed some simplification made by farmers in their decision-making process and offered suggestions for improving the decision support means to farmers. Both the models and the survey design are generic. They have been applied in a case study that makes it possible to explain the heterogeneity of fungicide use on soft wheat between farmers in the Midi-Pyrénées region by their personal characteristics, the characteristics of their production situation and their decision modes.

Nous étudions dans cette thèse la gestion des incertitudes au sein des systèmes à base de règles métier orientés objet (Object-Oriented Business Rules Management Systems ou OO-BRMS) et nous nous intersessions à des approches probabilistes. Afin de faciliter la modélisation des distributions de probabilités dans ces systèmes, nous proposons d'utiliser les modèles probabilistes relationnels (Probabilistic Relational Models ou PRM), qui sont une extension orientée objet des réseaux bayésiens. Lors de l'exploitation des OO-BRMS, les requêtes adressées aux PRM sont nombreuses et les réponses doivent être calculées rapidement. Pour cela, nous proposons, dans la première partie de cette thèse, un nouvel algorithme tirant parti de deux spécificités des OO-BRMS. Premièrement, les requêtes de ces derniers s'adressent seulement à une sous partie de leur base. Par conséquent, les probabilités à calculer ne concernent que des sous-ensembles de toutes les variables aléatoires des PRM. Deuxièmement, les requêtes successives diffèrent peu les unes des autres. Notre algorithme exploite ces deux spécificités afin d'optimiser les calculs. Nous prouvons mathématiquement que notre approche fournit des résultats exacts et montrons son efficacité par des résultats expérimentaux. Lors de la deuxième partie, nous établissons des principes généraux permettant d'étendre les OO-BRMS pour garantir une meilleure inter-operabilité avec les PRM. Nous appliquons ensuite notre approche au cas d'IBM Operational Decisions Manager (ODM) dans le cadre d'un prototype développé, que nous décrivons de manière générale. Enfin, nous présentons des techniques avancées permettant de compiler des expressions du langage technique d'ODM pour faciliter leur exploitation par le moteur probabiliste des PRM. / In this thesis, we address the issue of uncertainty in Object-Oriented Business Rules Management Systems (OO-BRMSs). To achieve this aim, we rely on Probabilistic Relational Models (PRMs). These are an object-oriented extension of Bayesian Networks that can be exploited to efficiently model probability distributions in OO-BRMSs. It turns out that queries in OO-BRMS are numerous and we need to request the PRM very frequently. The PRM should then provide a rapid answer. For this reason, we propose, in the first part of this thesis, a new algorithm that respects two specifities of OO-BRMSs and optimizes the probabilistic inference accordingly. First, OO-BRMSs queries affect only a subset of their base, hence, the probabilities of interest concern only a subset of the PRMs random variables. Second, successive requests differ only slightly from each other. We prove theoretically the correctness of the proposed algorithm and we highlight its efficiency through experimental tests. During the second part, we establish some principles for probabilistic OO-BRMSs and we describe an approach to couple them with PRMs. Then, we apply the approach to IBM Operational Decision Manager (ODM), one of the state-of-the-art OO-BRMSs, and we provide a general overview of the resulted prototype. Finally, we discuss advanced techniques to compile elements of ODM technical language into instructions that are exploitable by the PRM probabilistic engine.

Approved for public release; distribution in unlimited. / In a world in which the use of violence continues to be viewed as an acceptable method to pursue political goals, the use of terrorism as a political method will also continue. And within that world, in the coming years, hostage rescues and other direct actions to combat terrorism are likely to be the most frequent type of missions conducted by state special operations forces, including those of Romania. This thesis explores the importance of integrating three types of decisions-informational, structural, and operational-for the successful outcome of special operations. The thesis analyzes four operations, conducted by American, Belgian, and Israeli forces, and the circumstances of their positive or negative outcomes. The historical cases show that, if any one of the three types of decisions was not integrated with the other two, the operation was doomed. The analysis also reveals that the integration of decision-making can best be realized by using interagency coordination mechanisms and a collocation of decision-makers, especially for situations in which command arrangements are highly complicated or unclear. In light of the findings in the four case studies, an analysis of the Romanian Special Operations Forces reveals that its overall structure does not favor either immediate-response or high-complexity missions. The thesis concludes with a number of recommendations for short- and long-term mitigation of current command and control problems faced by Romanian Special Operations Forces. / Captain, Romanian Army